A few weeks ago, I accidentally got into a Twitter discussion of the serial, or Oxford, comma (this happens to me every couple months). A few programmer friends chimed in, and Dylan asked an interesting question: are programmers generally biased toward the Oxford comma? 18 minutes later, this survey happened.
I’m here today to share the results with you.
Given that this was a quick survey in response to a stupid Internet argument, I didn’t put much thought into the survey design. My questions were quick and to the point:
- How do you usually use the serial comma (also commonly known as the Oxford comma)? (Answers: Always, Never)
- Are you a programmer? (Answers: Yes, No)
A few people had problems with this.
The first question is somewhat flawed because it lets you choose only absolutes. I hoped this would be okay, for a few reasons:
- I included the word “usually” in the question
- The people who care enough to fill out this survey probably have strong enough feelings about the serial comma that one of the choices describes them pretty well
I got some feedback that the second question was confusing — did I only mean, “are you a professional programmer?” Or was a programming hobby sufficient? What if one was a programmer at a past job, but is no longer?
I figured that most people would know if they generally identified as a programmer-type-person.
Thanks to friends on Twitter and Facebook, I got 163 responses to the survey (after filtering out a bunch of prank duplicate submissions from a coworker — thanks, Mike!).
Of all 163 respondents:
- 137 (84.0%) said they Always use the comma. 26 (16.0%) said Never.
- 73 (44.8%) said they were programmers. 90 (55.2%) said they were not.
Of the 73 respondents who identified as programmers:
- 60 (82.2%) said they Always use the comma.
- 13 (17.8%) said Never.
Of the 90 non-programmer respondents:
- 77 (85.6%) said they Always use the comma.
- 13 (14.4%) said Never.
You can download the response data, minus IP addresses, here: serial_comma_anon.csv.
82.2% of programmer respondents said they always use the comma, compared to 85.6% of non-programmer respondents. These figures are equal enough to be within a reasonable margin of error with such a small sample size.
The data therefore suggest that programmers are more-or-less in line with the rest of the population in serial comma preference. This doesn’t support my original hypothesis (that programmers would be somewhat more inclined to use the comma consistently).